On Tue, May 06, 2003 at 10:07:14PM -0700, David S. Miller wrote:
> This doesn't even consider the case where the ipsec-tools copy of the
> headers becomes out of date with the kernel copy. This isn't a
> theoretical issue, this problem is real.
>
> For example, I just changed the values of a few SADB_EALG_* values in
> pfkeyv2.h. Now ipsec-tools is effectively broken. Oops, when will
> the copy in ipsec-tools get updated?
You just broke the userland ABI which must not happen. at all. That's
why userland having older headers is fine.
> What about applications, ie. normal ones, that want to pass IPSEC
> policies into the kernel via the socket options we have that allows
> per-socket IPSEC rules to be specified? The copy in ipsec-tools
> doesn't help them at all.
That's why we want the glibc-kernheader package. Or even better
a package of headers that can be used by the kernel and userland,
but this would require people to properly sort out kernel header
functionality like internal structures and prototypes/inlines from
the actual ABI-relevant contents. The networking headers currently
are very bad on this.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed May 07 2003 - 22:00:30 EST